Performance Of Model Research Articles

A Multimorbidity Analysis of Hospitalized Patients With COVID-19 in Northwest Italy: Longitudinal Study Using Evolutionary Machine Learning and Health Administrative Data.

Multimorbidity is a significant public health concern, characterized by the coexistence and interaction of multiple preexisting medical conditions. This complex condition has been associated with an increased risk of COVID-19. Individuals with multimorbidity who contract COVID-19 often face a significant reduction in life expectancy. The postpandemic period has also highlighted an increase in frailty, emphasizing the importance of integrating existing multimorbidity details into epidemiological risk assessments. Managing clinical data that include medical histories presents significant challenges, particularly due to the sparsity of data arising from the rarity of multimorbidity conditions. Also, the complex enumeration of combinatorial multimorbidity features introduces challenges associated with combinatorial explosions. This study aims to assess the severity of COVID-19 in individuals with multiple medical conditions, considering their demographic characteristics such as age and sex. We propose an evolutionary machine learning model designed to handle sparsity, analyzing preexisting multimorbidity profiles of patients hospitalized with COVID-19 based on their medical history. Our objective is to identify the optimal set of multimorbidity feature combinations strongly associated with COVID-19 severity. We also apply the Apriori algorithm to these evolutionarily derived predictive feature combinations to identify those with high support. We used data from 3 administrative sources in Piedmont, Italy, involving 12,793 individuals aged 45-74 years who tested positive for COVID-19 between February and May 2020. From their 5-year pre-COVID-19 medical histories, we extracted multimorbidity features, including drug prescriptions, disease diagnoses, sex, and age. Focusing on COVID-19 hospitalization, we segmented the data into 4 cohorts based on age and sex. Addressing data imbalance through random resampling, we compared various machine learning algorithms to identify the optimal classification model for our evolutionary approach. Using 5-fold cross-validation, we evaluated each model's performance. Our evolutionary algorithm, utilizing a deep learning classifier, generated prediction-based fitness scores to pinpoint multimorbidity combinations associated with COVID-19 hospitalization risk. Eventually, the Apriori algorithm was applied to identify frequent combinations with high support. We identified multimorbidity predictors associated with COVID-19 hospitalization, indicating more severe COVID-19 outcomes. Frequently occurring morbidity features in the final evolved combinations were age>53, R03BA (glucocorticoid inhalants), and N03AX (other antiepileptics) in cohort 1; A10BA (biguanide or metformin) and N02BE (anilides) in cohort 2; N02AX (other opioids) and M04AA (preparations inhibiting uric acid production) in cohort 3; and G04CA (Alpha-adrenoreceptor antagonists) in cohort 4. When combined with other multimorbidity features, even less prevalent medical conditions show associations with the outcome. This study provides insights beyond COVID-19, demonstrating how repurposed administrative data can be adapted and contribute to enhanced risk assessment for vulnerable populations.

Development and validation of a predictive model for PACU hypotension in elderly patients undergoing sedated gastrointestinal endoscopy.

Hypotension, characterized by abnormally low blood pressure, is a frequently observed adverse event in sedated gastrointestinal endoscopy procedures. Although the examination time is typically short, hypotension during and after gastroscopy procedures is frequently overlooked or remains undetected. This study aimed to construct a risk nomogram for post-anesthesia care unit (PACU) hypotension in elderly patients undergoing sedated gastrointestinal endoscopy. This study involved 2919 elderly patients who underwent sedated gastrointestinal endoscopy. A preoperative questionnaire was used to collect data on patient characteristics; intraoperative medication use and adverse events were also recorded. The primary objective of the study was to evaluate the risk of PACU hypotension in these patients. To achieve this, the least absolute shrinkage and selection operator (LASSO) regression analysis method was used to optimize variable selection, involving cyclic coordinate descent with tenfold cross-validation. Subsequently, multivariable logistic regression analysis was applied to build a predictive model using the selected predictors from the LASSO regression. A nomogram was visually developed based on these variables. To validate the model, a calibration plot, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA) were used. Additionally, external validation was conducted to further assess the model's performance. The LASSO regression analysis identified predictors associated with an increased risk of adverse events during surgery: age, duration of preoperative water abstinence, intraoperative mean arterial pressure (MAP) <65mmHg, decreased systolic blood pressure (SBP), and use of norepinephrine (NE). The constructed model based on these predictors demonstrated moderate predictive ability, with an area under the ROC curve of 0.710 in the training set and 0.778 in the validation set. The DCA indicated that the nomogram had clinical applicability when the risk threshold ranged between 20 and 82%, which was subsequently confirmed in the external validation with a range of 18-92%. Incorporating factors such as age, duration of preoperative water abstinence, intraoperative MAP <65mmHg, decreased SBP, and use of NE in the risk nomogram increased its usefulness for predicting PACU hypotension risk in elderly patient undergoing sedated gastrointestinal endoscopy.

Evaluation of artificial intelligence-powered screening for sexually transmitted infections-related skin lesions using clinical images and metadata

BackgroundSexually transmitted infections (STIs) pose a significant global public health challenge. Early diagnosis and treatment reduce STI transmission, but rely on recognising symptoms and care-seeking behaviour of the individual. Digital health software that distinguishes STI skin conditions could improve health-seeking behaviour. We developed and evaluated a deep learning model to differentiate STIs from non-STIs based on clinical images and symptoms.MethodsWe used 4913 clinical images of genital lesions and metadata from the Melbourne Sexual Health Centre collected during 2010–2023. We developed two binary classification models to distinguish STIs from non-STIs: (1) a convolutional neural network (CNN) using images only and (2) an integrated model combining both CNN and fully connected neural network (FCN) using images and metadata. We evaluated the model performance by the area under the ROC curve (AUC) and assessed metadata contributions to the Image-only model.ResultsOur study included 1583 STI and 3330 non-STI images. Common STI diagnoses were syphilis (34.6%), genital warts (24.5%) and herpes (19.4%), while most non-STIs (80.3%) were conditions such as dermatitis, lichen sclerosis and balanitis. In both STI and non-STI groups, the most frequently observed groups were 25–34 years (48.6% and 38.2%, respectively) and heterosexual males (60.3% and 45.9%, respectively). The Image-only model showed a reasonable performance with an AUC of 0.859 (SD 0.013). The Image + Metadata model achieved a significantly higher AUC of 0.893 (SD 0.018) compared to the Image-only model (p < 0.01). Out of 21 metadata, the integration of demographic and dermatological metadata led to the most significant improvement in model performance, increasing AUC by 6.7% compared to the baseline Image-only model.ConclusionsThe Image + Metadata model outperformed the Image-only model in distinguishing STIs from other skin conditions. Using it as a screening tool in a clinical setting may require further development and evaluation with larger datasets.

Risk assessment tool for anemia of chronic disease in systemic lupus erythematosus: a prediction model.

This study aims to develop a predictive model for estimating the likelihood of anemia of chronic disease (ACD) in patients with systemic lupus erythematosus (SLE) and to elucidate the relationship between various factors and ACD METHODS: Individuals diagnosed with SLE for at least one year were enrolled and categorized into two groups: those with ACD and those without anemia symptoms. Patients were randomly assigned to training and test sets at an 8:2 ratio. The least absolute shrinkage and selection operator (LASSO) method was used to select predictors, followed by logistic regression for modeling. Model performance was assessed using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) for both training and test sets. The study included a total of 216 patients, with 172 in the training set and 44 in the test set. LASSO identified 6 variables for constructing the predictive model, resulting in an area under the curve (AUC) of 0.833 (95% CI, 0.773-0.892) in the training set and 0.861 (95% CI, 0.750-0.972) in the test set. Calibration curves indicated consistency between expected and observed probabilities. DCA indicated that the model yielded a net benefit with threshold probabilities ranging from 20% to 90% in the training set and from 10% to 80% in the test set. This study presents a predictive model for assessing the risk of ACD in SLE patients. The model effectively captures the underlying mechanism of ACD in SLE and empowers clinicians to make well-informed treatment adjustments. Key Points • Development of a New Predictive Model: This study introduces a new predictive model to evaluate the likelihood of anemia of chronic disease (ACD) in patients with systemic lupus erythematosus (SLE). The model utilizes routine laboratory parameters to identify high-risk individuals, addressing a significant gap in current clinical practice. • Reflection of Potential Mechanisms for ACD Development: By incorporating the factors needed to construct the predictive model, this study also sheds light on the potential mechanisms of ACD development in SLE patients.

Predictors of Symptom Reduction and Remission Among People with Anxiety: Secondary Analyses from a Randomized Controlled Trial.

Despite the substantial disease burden of anxiety disorders, only limited or conflicting data on prognostic factors is available. Most studies include patients in the secondary healthcare sector thus, the generalizability of findings is limited. The present study examines predictors of symptom reduction and remission in patients with anxiety disorders in a primary care setting. 214 patients with anxiety disorders, recruited as part of the Collabri Flex trial, were included in secondary analyses. Data on potential predictors of anxiety symptoms at 6-month follow-up was collected at baseline, including patient characteristics related to demography, illness, comorbidity, functional level, life quality, and self-efficacy. The outcomes were symptom reduction and remission. Univariate and multivariate linear and logistic regression analyses were conducted to assess the associations between predictor variables and the outcome, and machine-learning methods were also applied. In multiple linear regression analysis, anxiety severity at baseline (β = -6.05, 95% CI = -7.54,-4.56, p < 0.001) and general psychological problems and symptoms of psychopathology (SCL-90-R score) (β = 2.19, 95% CI = 0.24,4.14, p = 0.028) were significantly associated with symptom change at 6months. Moreover, self-efficacy was associated with the outcome, however no longer significant in the multiple regression model. In multiple logistic regression analysis, anxiety severity at baseline (OR = 0.54, 95% CI = -1.13,-0.12, p = 0.018) was significantly associated with remission at 6months. There was no predictive performance of the machine-learning models. Our study contributes with information that could be valuable knowledge for managing anxiety disorders in primary care.

Assessment of Arctic sea ice simulations in cGENIE model and projections under RCP scenarios

Simulating and predicting Arctic sea ice accurately remains an academic focus due to the complex and unclear mechanisms of Arctic sea ice variability and model biases. Meanwhile, the relevant forecasting and monitoring authorities are searching for models to meet practical needs. Given the previous ideal performance of cGENIE model in other fields and notable features, we evaluated the model’s skill in simulating Arctic sea ice using multiple methods and it demonstrates great potential and combined advantages. On this basis, we examined the direct drivers of sea-ice variability and predicted the future spatio-temporal changes of Arctic sea ice using the model under different Representative Concentration Pathways (RCP) scenarios. Further studies also found that Arctic sea ice concentration shows large regional differences under RCP 8.5, while the magnitude of the reduction in Arctic sea ice thickness is generally greater compared to concentration, showing a more uniform consistency of change.

Intelligent diagnostic model for pterygium by combining attention mechanism and MobileNetV2

AIM: To evaluate the application of an intelligent diagnostic model for pterygium. METHODS: For intelligent diagnosis of pterygium, the attention mechanisms—SENet, ECANet, CBAM, and Self-Attention—were fused with the lightweight MobileNetV2 model structure to construct a tri-classification model. The study used 1220 images of three types of anterior ocular segments of the pterygium provided by the Eye Hospital of Nanjing Medical University. Conventional classification models—VGG16, ResNet50, MobileNetV2, and EfficientNetB7—were trained on the same dataset for comparison. To evaluate model performance in terms of accuracy, Kappa value, test time, sensitivity, specificity, the area under curve (AUC), and visual heat map, 470 test images of the anterior segment of the pterygium were used. RESULTS: The accuracy of the MobileNetV2+Self-Attention model with 281 MB in model size was 92.77%, and the Kappa value of the model was 88.92%. The testing time using the model was 9ms/image in the server and 138ms/image in the local computer. The sensitivity, specificity, and AUC for the diagnosis of pterygium using normal anterior segment images were 99.47%, 100%, and 100%, respectively; using anterior segment images in the observation period were 88.30%, 95.32%, and 96.70%, respectively; and using the anterior segment images in the surgery period were 88.18%, 94.44%, and 97.30%, respectively. CONCLUSION: The developed model is lightweight and can be used not only for detection but also for assessing the severity of pterygium.

Detection and severity assessment of obstructive sleep apnea according to deep learning of single-lead electrocardiogram signals.

Developing a convenient detection method is important for diagnosing and treating obstructive sleep apnea. Considering availability and medical reliability, we established a deep-learning model that uses single-lead electrocardiogram signals for obstructive sleep apnea detection and severity assessment. The detection model consisted of signal preprocessing, feature extraction, time-frequency domain information fusion, and classification segments. A total of 375 patients who underwent polysomnography were included. The single-lead electrocardiogram signals obtained by polysomnography were used to train, validate and test the model. Moreover, the proposed model performance on a public dataset was compared with the findings of previous studies. In the test set, the accuracy of per-segment and per-recording detection were 82.55% and 85.33%, respectively. The accuracy values for mild, moderate and severe obstructive sleep apnea were 69.33%, 74.67% and 85.33%, respectively. In the public dataset, the accuracy of per-segment detection was 91.66%. A Bland-Altman plot revealed the consistency of true apnea-hypopnea index and predicted apnea-hypopnea index. We confirmed the feasibility of single-lead electrocardiogram signals and deep-learning model for obstructive sleep apnea detection and severity evaluation in both hospital and public datasets. The detection performance is high for patients with obstructive sleep apnea, especially those with severe obstructive sleep apnea.

Evaluation of transfer ensemble learning-based convolutional neural network models for the identification of chronic gingivitis from oral photographs

BackgroundTo evaluate the performances of several advanced deep convolutional neural network models (AlexNet, VGG, GoogLeNet, ResNet) based on ensemble learning for recognizing chronic gingivitis from screening oral images.MethodsA total of 683 intraoral clinical images acquired from 134 volunteers were used to construct the database and evaluate the models. Four deep ConvNet models were developed using ensemble learning and outperformed a single model. The performances of the different models were evaluated by comparing the accuracy and sensitivity for recognizing the existence of gingivitis from intraoral images.ResultsThe ResNet model achieved an area under the curve (AUC) value of 97%, while the AUC values for the GoogLeNet, AlexNet, and VGG models were 94%, 92%, and 89%, respectively. Although the ResNet and GoogLeNet models performed best in classifying gingivitis from images, the sensitivity outcomes were not significantly different among the ResNet, GoogLeNet, and Alexnet models (p>0.05). However, the sensitivity of the VGGNet model differed significantly from those of the other models (p < 0.001).ConclusionThe ResNet and GoogLeNet models show promise for identifying chronic gingivitis from images. These models can help doctors diagnose periodontal diseases efficiently or based on self-examination of the oral cavity by patients.

Development of a nomogram to predict recurrence scores obtained using Oncotype DX in Japanese patients with breast cancer.

Chemotherapy is crucial for hormone receptor-positive, human epidermal growth factor receptor 2 (HER2)-negative breast cancer, and its survival benefits may outweigh adverse events. Oncotype DX (ODX) assesses this balance; however, it is expensive. Using nomograms to identify cases requiring ODX may be economically beneficial. We aimed to identify clinicopathological variables that correlated with the recurrence score (RS) and develop a nomogram that predicted the RS. We included 457 patients with estrogen receptor-positive, HER2-negative breast cancer with metastases in fewer than four axillary lymph nodes who underwent surgery and ODX at our hospital between 2007 and 2023. We developed nomograms and internally validated them in 310 patients who underwent surgery between 2007 and 2021 and validated the model's performance in 147 patients who underwent surgery between 2022 and 2023. Logistic regression analysis revealed that progesterone receptor (PgR) level, histological grade (HG), and Ki67 index independently predicted the RS. A nomogram was developed using these variables to predict the RS (area under the curve [AUC], 0.870; 95% confidence interval [CI], 0.82-0.92). The nomogram was applied to the model validation group (AUC, 0.877; 95% CI, 0.80-0.95). When the sensitivity of the nomogram was 90%, the model was able to identify 52.3% low-RS and 41.2% high-RS cases not requiring ODX. This was the first nomogram model developed based on data from a cohort of Japanese women. It may help determine the indications for ODX and the use of nomogram to identify cases requiring ODX may be economically beneficial.

A novel structure grey prediction model with strong compatibility and its application in forecasting the annual average concentration of particulate matter in Beijing

Beijing has long suffered from the events of fine particulate matter pollution. Consequently, it is of paramount importance to be able to scientifically and reasonably predict the concentration of particulate matter. To this end, a novel prediction model with an adaptive grey generation operator and fractional derivative is proposed. The novel model initially expands the range values for the order of the grey generation operator to enhance data preprocessing capabilities. Meanwhile, the introduction of fractional derivatives enhances the model’s adaptability and overcomes the limitations of traditional grey prediction models that use integer-order differential equations. Subsequently, the fractional derivative and operator orders of the novel model are optimized to find the best parameter combination, thereby enhancing the model's performance. Furthermore, the novel model is compatible with seven other models when its parameters take different values, thus expanding the model’s applicability. The novel model is then applied to simulate and predict the particulate matter concentration in Beijing, achieving a comprehensive mean relative percentage error of only 1.86%, significantly outperforming several similar models(17.56%,17.64%,3.42%). Finally, the paper presents a forecast of the particulate matter concentration in Beijing over the next four years.

The incremental design of a machine learning framework for medical records processing.

This work presents the development and evaluation of coordn8, a web-based application that streamlines fax processing in outpatient clinics using a "human-in-the-loop" machine learning framework. We demonstrate the effectiveness of the platform at reducing fax processing time and producing accurate machine learning inferences across the tasks of patient identification, document classification, spam classification, and duplicate document detection. We deployed coordn8 in 11 outpatient clinics and conducted a time savings analysis by observing users and measuring fax processing event logs. We used statistical methods to evaluate the machine learning components across different datasets to show generalizability. We conducted a time series analysis to show variations in model performance as new clinics were onboarded and to demonstrate our approach to mitigating model drift. Our observation analysis showed a mean reduction in individual fax processing time by 147.5 s, while our event log analysis of over 7000 faxes reinforced this finding. Document classification produced an accuracy of 81.6%, patient identification produced an accuracy of 83.7%, spam classification produced an accuracy of 98.4%, and duplicate document detection produced a precision of 81.0%. Retraining document classification increased accuracy by 10.2%. coordn8 significantly decreased fax-processing time and produced accurate machine learning inferences. Our human-in-the-loop framework facilitated the collection of high-quality data necessary for model training. Expanding to new clinics correlated with performance decline, which was mitigated through model retraining. Our framework for automating clinical tasks with machine learning offers a template for health systems looking to implement similar technologies.

Multi-modal fusion and feature enhancement U-Net coupling with stem cell niches proximity estimation for voxel-wise GBM recurrence prediction.

We aim to develop a Multi-modal Fusion and Feature Enhancement U-Net (MFFE U-Net) coupling with stem cell niche (SCN) proximity estimation to improve voxel-wise GBM recurrence prediction.&#xD;Approach: 57 patients with pre- and post-surgery MR scans were retrospectively solicited from 4 databases. Post-surgery MR scans included two months before the clinical diagnosis of recurrence and the day of the radiologically confirmed recurrence. The recurrences were manually annotated on the T1ce. The high-risk recurrence (HRR) region was first determined. Then, a sparse multi-modal feature fusion U-Net was developed. The 50 patients from 3 databases were divided into 70% training, 10% validation, and 20% testing. 7 patients from the 4th institution were used as external testing with transfer learning. Model performance was evaluated by recall, precision, F1-score, and Hausdorff Distance at the 95% percentile (HD95). The proposed MFFE U-Net was compared to the support vector machine (SVM) model and two state-of-the-art neural networks. An ablation study was performed.&#xD;Main results: The MFFE U-Net achieved a precision of 0.79±0.08, a recall of 0.85±0.11, and an F1-score of 0.82±0.09. Statistically significant improvement was observed when comparing MFFE U-Net with proximity estimation couple SVM (SVMPE), mU-Net, and Deeplabv3. The HD95 was 2.75±0.44mm and 3.91±0.83mm for the 10 patients used in the model construction and 7 patients used for external testing, respectively. The ablation test showed that all five MR sequences contributed to the performance of the final model, with T1ce contributing the most. Convergence analysis, time efficiency analysis, and visualization of the intermediate results further discovered the characteristics of the proposed method.&#xD;Significance: We present an advanced multi-modal fusion and feature enhancement learning framework, MFFE U-Net, for effective voxel-wise GBM recurrence prediction. MFFE U-Net performs significantly better than the state-of-the-art networks and can potentially guide early RT intervention of the disease recurrence.

Performance of risk prediction models for diabetic foot ulcer: a meta-analysis

Performance of risk prediction models for diabetic foot ulcer: a meta-analysis

Optimizing data acquisition: a Bayesian approach for efficient machine learning model training

Acquiring a substantial number of data points for training accurate machine learning (ML) models is a big challenge in scientific fields where data collection is resource-intensive. Here, we propose a novel approach for constructing a minimal yet highly informative database for training ML models in complex multi-dimensional parameter spaces. To achieve this, we mimic the underlying relation between the output and input parameters using Gaussian process regression (GPR). Using a set of known data, GPR provides predictive means and standard deviation for the unknown data. Given the predicted standard deviation by GPR, we select data points using Bayesian optimization to obtain an efficient database for training ML models. We compare the performance of ML models trained on databases obtained through this method, with databases obtained using traditional approaches. Our results demonstrate that the ML models trained on the database obtained using Bayesian optimization approach consistently outperform the other two databases, achieving high accuracy with a significantly smaller number of data points. Our work contributes to the resource-efficient collection of data in high-dimensional complex parameter spaces, to achieve high precision ML predictions.

Policy Learning for Actively Labeled Sample Selection on Lumbar Semi-supervised Classification.

Large labeled data bring significant performance improvement, but acquiring labeled medical data is particularly challenging due to the laborious, time-consuming, and medically qualified annotation. Semi-supervised learning has been employed to leverage unlabeled data. However, the quality and quantity of annotated data have a great influence on the performance of the semi-supervised model. Selecting informative samples through active learning is crucial and could improve model performance. Therefore, we propose a unified semi-supervised active learning architecture (RL-based SSAL) that alternately trains a semi-supervised network and performs active sample selection. Semi-supervised model is first well trained for sample selection, and selected label-required samples are annotated and added to the previously labeled dataset for subsequent semi-supervised model training. To learn to select the most informative samples, we adopt a policy learning-based approach that treats sample selection as a decision-making process. A novel reward function based on the product of predictive confidence and uncertainty is designed, aiming to select samples with high confidence and uncertainty. Comparisons with a semi-supervised baseline on collected lumbar disc herniation dataset demonstrate the effectiveness of the proposed RL-based SSAL, achieving over 3% promotion across different amounts of labeled data. Comparisons with other active learning methods and ablation studies reveal the superiority of proposed policy learning based on active sample selection and reward function. Our model trained with only 200 labeled data achieves an accuracy of 89.32% which is comparable to the performance achieved with the entire labeled dataset, demonstrating its significant advantage.

Systematic Evaluation of Hematoma Expansion Models in Spontaneous Intracerebral Hemorrhage: A Meta-Analysis and Meta-Regression Approach.

Accurate prediction of hematoma expansion (HE) in spontaneous intracerebral hemorrhage (sICH) is crucial for tailoring patient-specific treatments and improving outcomes. Recent advancements have yielded numerous HE risk factors and predictive models. This study aims to evaluate the characteristics and efficacy of existing HE prediction models, offering insights for performance enhancement. A comprehensive search was conducted in PubMed for observational studies and randomized controlled trials focusing on HE prediction, written in English. The prediction models were categorized based on their incorporated features and modeling methodology. Rigorous quality and bias assessments were performed. A meta-analysis of studies reporting C-statistics was executed to assess and compare the performance of current HE prediction models. Meta-regression was utilized to explore heterogeneity sources. From 358 initial records, 22 studies were deemed eligible, encompassing traditional models, hematoma imaging feature models, and models based on artificial intelligence (AI) or radiomics. Meta-analysis of 11 studies, involving 12,087 sICH patients, revealed an aggregated C-statistic of 0.74 (95% CI: 0.69 - 0.78) across seven HE prediction models. Eight characteristics related to development cohorts were identified as key factors contributing to performance variability among these models. The findings indicate that the current predictive capacity for HE risk remains suboptimal. Enhanced accuracy in HE prediction is vital for effectively targeting patient populations most likely to benefit from tailored treatment strategies.

Thermal Performance of Cattaneo-Christov Heat Flux Model on 3D Micropolar Fluid with Motile Microbes Across a Permeable Surface

Thermal Performance of Cattaneo-Christov Heat Flux Model on 3D Micropolar Fluid with Motile Microbes Across a Permeable Surface

BagFormer: Better cross-modal retrieval via bag-wise interaction

In the field of cross-modal retrieval, single encoder models tend to perform better than dual encoder models, but they suffer from high latency and low throughput. In this paper, we propose a dual encoder model called BagFormer that utilizes bag-wise late interaction mechanism to improve re-rank performance without sacrificing latency and throughput. BagFormer achieves this by employing a bagging layer, which facilitates the transformation of text to an appropriate granularity. This not only mitigates the issue of modal granularity mismatch but also enables the integration of entity knowledge into the model. Our experiments have shown that BagFormerViT-B outperforms the traditional dual-encoder model CLIPViT-B by 7.97% in zero-shot settings. Under fine-tuned conditions, BagFormerViT-B demonstrates an even more significant improvement of 17.98% over CLIPViT-B. Moreover, BagFormer not only matches the performance of cutting-edge single-encoder models in cross-modal retrieval tasks but also provides efficient inference processes characterized by lower latency and higher throughput. Compared to single-encoder models, BagFormer can achieve a speedup ratio of 38.14 when re-ranking individual candidates. Code and models are available at github.com/howard-hou/BagFormer.

Optimising synthetic datasets for machine learning-based prediction of building damage due to tunnelling

Assessment of tunnelling-induced building damage is a complex Soil-Structure Interaction (SSI) probelm, influenced by numerous geometric and material parameters of both the soil and structures, and is characterised by strong non-linear behaviour. Currently, there is a trend towards developing data-driven models using Machine Learning (ML) to capture this complex behaviour. Given the scarcity of real data, which typically comes from specific case studies, many researchers have turned to creating extensive synthetic datasets via sophisticated and validated numerical models like Finite Element Method (FEM). However, the development of these datasets and the training of advanced ML algorithms present significant challenges. poses significant challenges. Reliance solely on parameter domains and ranges derived from case studies can lead to imbalanced data distributions and subsequently poor performance of models in less populated regions. In this paper, we introduce a strategy for designing optimal high-confidence datasets through an iterative procedure. This process begins with a systematic literature review to determine the importance of parameters, their ranges, and dependencies as they pertain to building damage induced by SSI. Starting with several hundred FEM simulations, we generate an initial dataset and assess its quality and impact through Sensitivity Analysis (SA) studies, statistical modelling, and re-sampling in statistically significant regions. This evaluation allows us to refine the model’s input space, seeking scenarios that mitigate output distribution imbalances. The procedure is repeated until the datasets achieve a satisfactory balance for training metamodels, minimising bias effectively. Our findings highlight the success of this approach in identifying an optimal and feasible input space that significantly reduces imbalanced distributions of output features. This approach not only proves effective in our study but also offers a versatile methodology that could be adapted to other disciplines aiming to generate high-quality synthetic datasets.